Polarization and Readout Electronics of the Ionization Channel

1+Q∆V 2.97

(4.24) The Luke-Neganov effect is a very interesting strategy to enhance the sensitivity of the heat channel. However, this heat channel boost is applied at the expanse of the discrimination ability of the detectors thanks to the double heat-ionization energy measurement. When operating the detector with voltages higher than∼10 V, the Luke-Neganov boost dominates in the expression of the measured heat energyEheat:

∆V≫ ^2.97

Q ⇒ Eheat ≈ENL∝EIon. (4.25)

The heat energy becomes equals to the ionization energy ignoring the multiplicative factor. As such, it becomes impossible to distinguish discriminate the electronic recoils from the nuclear recoils as their signature in the same. In fact, this observation can be further at any voltage bias:

the existence of the Luke-Neganov effect diminishes the discrimination power of the ER and NR events. As this feature is essential for the correct operation of the detector and to reach their scientific goals, the Luke-Neganov effect should be kept minimal by aiming at low voltage bias in the detector design (e.g.∆V∼8 V for FID800 detectors of EDELWEISS-III).

4.1.8 Polarization and Readout Electronics of the Ionization Channel

The electronics associated with the ionization channel has two functions: the signal readout and the polarization of the electrodes. As usual, the electronics are separated in two stages: the cold electronic inside the cryostat, before any signal amplification, and the hot electronics outside of the cryostat transferring the amplified signal. This section focuses on the cold electronics which has major influence on the resolution of the ionization channel and the operation of the ioniza-tion channel. The readout and the polarizaioniza-tion of all the electrodes of a detector are driven by a single computer. However, we can consider that each electrode is equipped with an independent cold electronic line.

Figure 4.5: Scheme of a cold electronics line for the ionization channel currently used at IP2I (left) and in development with HEMT technology (right). The electronics line assures the polarization of the electrode and the signal readout. One of the main difference is the floating state of the current electronic (heaviside signal) versus the polarization resistance of the HEMT (decaying exponential signal).

Figure4.5shows two electric circuits of the electronic line of a single electrode of a detector.

The circuit on the left represents the current electronics of the IP2I cryostat used to realize all the ionization measurements presented in this work. It uses Junction Field Effect Transistor (JFET).

On the right is a representation of a High Electron Mobility Transistor (HEMT) based electronics in development for building the RICOCHETexperiment [35] and enhancing the performance of the EDELWEISSexperiment.

The readout of the JFET-based and HEMT-based electronics lines measure the same physical quantity: the electric potential of the electrode. The JFET and HEMT are used to readout and amplify the voltage signal. This method can differ from other experiments equipped for reading the current induced by the drifting charge on a feedback capacitance with a charge amplifier.

Compared to a charge amplifier, a voltage amplifier does not involve any resistor in the ampli-fication scheme, resulting a lower electronic noise. However, the use of a voltage amplifier is only possible with low leakage current lower than 0.1 fA. One should note that the amplifying transistor situated at the end of the electronic line is separated from the electrode, represented by the capacitanceC_d, by a coupling capacitor of higher value Cc ∼ O(1 nF). The role of this decoupling capacitanceCd is to isolate the transistor gate from the bias voltage applied on the electrodes of the detector.

The electronics lines are designed to assure their functions while keeping the electronics noise as low as possible. Once the signal is amplified, it is transmitted to the acquisition com-puter through the hot electronics without adding additional noise.

There are several differences between the current JFET-based and the future HEMT-based electronics. The main change concerns the technology of the voltage amplifier. The JFET are operated at a low temperature of 100 K inside the cryostat. At the time of design for the EDW-III experiment, this technology offered the best performances. Nowadays, recent breakthrough in the electronics field adapts the HEMT technology to the constraints of the EDELWEISS and RICOCHET experiments. With a HEMT operating at the lower cryogenic temperature of 1 K, the thermal noise of the electronic components is reduced leading to an overall reduced noise injection in the amplification scheme.

The other main differences resides in the polarization of the electrodes. The polarization is the process of fixing the electric potential of an electrode. In the current JFET-based electronics, this process is assured by a mechanical relay (represented by the polarization switch on the fig-ure4.5) linking the electrode to a constant voltage source. By switching on, the electrode, acting as the plate of a capacitor, is accumulating charges and eventually reaches the aimed poten-tial. By switching off, the electrode keeps the accumulating charge and its now at fixed electric potential. Due to the relay opening the circuit, the constant voltage source cannot induce any

electronic noise on the electrode. The downside of this method is the progressive neutraliza-tion of the polarizaneutraliza-tion as the electrode collects charges. Moreover, in the case of the RICOCHET

experiment with surface operation of the bolometers, the event rate is high leading to a faster discharge. The counter is to periodically switching on and off the relay in order to re-establish the wanted potential on the electrode. This procedure is called “maintenance”. In reality, a main-tenance lasts one minute and consists in multiples relay switches and relay changeovers. During this one minute of maintenance, the detector is not available for data taking. This is enforced by the "maintenance cut" described in section 6.2.5. The frequency of maintenance is adapted to the condition of operation of the detector: too much maintenance induces unnecessary dead time while too few leads to a lower and uncontrolled voltage bias. For above ground operation at IP2I, the usual maintenance period is of about 30 minutes and should be empirically adjusted to the detector and the event rate. For the HEMT-based electronics, the polarization is contin-uously assured by a constant voltage sourceeDAQ in series with a resistor of high impedance R_Polar. = 10¹⁰Ω. While this method induced supplementary electronic noise, there is no more dead time and the maintenance procedure is not needed.

The ionization channel readout is sampled at an initial frequency of 100 kHz. This means the period of the measurement points is 10 µs which is still greater than the estimated time span of an ionization signal of a few µs. As a result, an ionization signal is recorded as an Heaviside function. No information can be obtained on the shape of the signal. The high readout sampling was historically chosen in EDELWEISS for the purpose of synchronizing an external detector vetoing incoming muons with a good timing resolution. The highly sampled ionization signal is then averaged in order to produce a signal of sampling frequency fs. This sampling frequency is a parameter of the acquisition and can be set in the range [200, 1000] Hz. This lower sampling records the information contained in the Heaviside-shaped ionization signal (its amplitude mainly) while being lightweight in term of disk space, which is essential considering the recording and processing resources at our disposal. The saved ionization signal of sampling frequency fsis composed of points whose values are averages of 100 kHz/fspoints. In the end, the saved signal of an event of sampling frequency fsis therefore Heaviside-like function H(t) with a varying valueH(t=0).

The ionization signal is readout as a digital signal which can take, by definition, a limited number of values. This readout range contains 64000 values which are sorted on the so-called readout dynamic range [−^32000,+32000]expressed in Analog-to-Digital Unit (ADU). For the ionization channel, the correspondence between ADU and Volt is 67.4 nV/ADU. As such, the dynamic range can records perturbation with a maximum amplitude of 2.156 V. As a reference, the sensitivity of the ionization channel is about 60ADU/keV ≈ ^{4 µV}·^keV−1. A high event rate of low energy events or single high energy events can rapidly make the data stream fall out of this dynamic range, effectively saturating the acquisition electronics. As to avoid this phenomenon and correct it, both electronics circuits feature a feedback line. A procedure called

“reset” consists in linking the input of the amplifying transistor to a null potential as to recenter the recorded signal on 0 ADU. The linking is assured in the feedback line with mechanical relays.

Each reset induces an artifact signal on both the heat and ionization channels. While it can be easily discriminate from real events due to its shape, these artifact signals are discarded by applying the reset cut during the data analysis presented in the section 6.2.5. As such, reset procedures creates dead times during which the detector is not available for data acquisition.

Similarly to the maintenance period, the period of the reset is empirically adjusted to reach a balance between dead time and dynamic range saturation. The period of the resets is of a few seconds. In the case of surfaces operation at IP2I, the event rate is high and lot of charges are accumulated which needs for a shorter period before reset than an underground operation.